Main
James Gammerman
Education
MSc, Machine Learning
Royal Holloway University
London, UK
2018 - 2016
- Grade: Distinction
- Selected modules: Deep Learning, Tree-based methods, Regression, Unsupervised Learning, Natural language processing, Large-scale Data Analytics (Big Data)
- Thesis: Predictive Maintenance with Conformal and Probabilistic Prediction - A Commercial Case Study. Won best thesis award.
MSci, Chemistry
Imperial College
London, UK
2013 - 2009
- Grade: 2:1
- Masters project in computational chemical physics: Heat Transfer in Ionic Liquids
Professional Experience
Data Scientist
Centrica
London, UK
Current - 2017
- Responsible for data exploration and cleaning, feature engineering, modelling and reporting/deployment
- Participated in several value-adding projects related to predictive maintenance, revenue forecasting and safety incidents
- Gained experience with modern ML frameworks such as Tensorflow and Apache Spark and cloud architectures such as AWS and Microsoft Azure
- Also responsible for some employee training and recruitment (see below)
Business Analyst
ExxonMobil
London, UK
2017 - 2013
- Numerous analytical tasks relating to European gas and power markets
Academic Publications
Journal article: Multi-level conformal clustering: A distribution-free technique for clustering and anomaly detection
Neurocomputing, Volume 397, 2020, pp. 279-291
N/A
2020
- Introduces a novel clustering technique which also incorporates anomaly detection
Poster abstract: Conformal Anomaly Detection based on Association Rules
Proceedings of Machine Learning Research, Volume 105, 2019, pp.246-7
N/A
2019
- Introduces a novel machine learning technique for database cleaning
Talks and Teaching
Workshop: Business Analytics in R - An Introduction to Statistical Programming
Centrica
Various Locations
Current - 2018
- I have presented this workshop several times to company employees across the UK.
Talk: Machine Learning: Progress & Prospects
Odessa (Mechnikov) University
Odessa, Ukraine
2018
- Guest lecture at Data Science meetup; available on Youtube.